13 notes on user research
For a customer-centric team
Research is a useful approach to gain conviction for your problem or solution. It gives you qualitative signal. Here’s a few notes:
1. There are different levels of qualitative evidence. From the weakest to strongest:
Users say something
Users say something they've done in the past
Users tell specific stories about what they've done
Users show you what they do
You observe the users
2. If you have an opportunity to observe users in their natural environment, you should invest your time in that.
3. Natural environment: You go to the user's place. Lab environment: You bring users to your office or zoom.
4. Product Managers shouldn't outsource the research. Product Managers should involve themselves in the research activity, and make the time. The amount of context and nuance will help you sharpen your product intuition and strategy.
5. The goal of the user interview is to understand the mental model and how the users operate. The user research is not to ask what solution users want. Stripe’s CEO has shared the similar view on his Tweet.
6. When I'm discovering the problem, I look for 5 things when interviewing users:
7. Before your user interview, prepare the debrief questions. After your interview, you should know to have at least a few reflection questions as a team. "Do we see users can accomplish X?" "Can users easily understand this content?"
8. Always debrief after your interview. Don't skip it. Spend at least 15 minutes. In that debrief, your team should understand what to adjust for the next interview as well.
9. If the key people can join the research activity, you don't have to spend time preparing the research report. After debriefing, you can just have a simple document with a bullet list to capture the key insights.
10. Don't ask for solutions from users. You should find and dig deeper into why they want the solution—then, you might discover the real problem or needs. Don’t ask what feature the users want. It’s not their job to figure out the next feature to build because people are terrible at predicting that.
11. Don't ask a hypothetical question like, "Would you buy this product?" You'll get bad data.
12. You can't validate everything from the qualitative. Sometimes, you need to launch in order to really understand if people actually use the product. Remember to launch with a minimum effort first.
13. But, launching without any qualitative validation or signal is actually a waste of time. You can at least get a signal: - Is the interaction good? - Is the content good? - Do people get excited about this? - Is our product actually better than the competitor's?